We employ computer simulltion and mathematical analysis to meet anatomical an neurophysiological constraints in understanding how observed structures may help subserve observable functions, building models in tight interaction with experimentalists. Two areas serve as foci for the research: 1. Learning and development in neural systems - We build models to explore the possible roles of synaptic modification, selective preservation of symapses, ans axonal competition in cortical maturation and in the formation of retino-tectal connections. Perception and control of movement -The role of optic flow analysis in locomotion; models of spinal cord and cerebellum; segmentation of visual input on depth and motion cues; and neural mechanisms of visually guided behavior in amphibians and primates. These efforts contibute to development of a general theoretical framework which includes the use of stability techniques from differential equation theory to study competition and cooperation in layered neural networks.